Water conservancy knowledge graph dynamic real-time updating system and method based on physical-information depth coupling and semantic field intensity driving

Abstract

The invention relates to the technical field of water conservancy knowledge graph updating, and discloses a water conservancy knowledge graph dynamic real-time updating system and method based on physical-information deep coupling and semantic field intensity driving, and the system comprises a physical sensing layer which is used for carrying out the real-time collection, transmission and preprocessing of multi-source water conservancy physical data; the coupling mapping layer is used for mapping the multi-source water conservancy physical data into semantic state variables available for a knowledge graph; the field intensity calculation layer is used for calculating semantic field intensity based on semantic state variable changes and determining nodes and relations, needing to be updated, of the knowledge graph; the graph operation layer is used for executing knowledge graph updating decision making and writing operation according to a semantic field intensity triggering threshold value; and the knowledge service layer is used for providing dynamic knowledge service for an external intelligent water conservancy system. According to the invention, multi-source water conservancy data can be effectively integrated, comprehensive acquisition and preprocessing of real-time data are realized through the physical sensing layer, and the accuracy and integrity of the data are ensured.

Description

A dynamic real-time update system and method for water conservancy knowledge graph based on deep physical-information coupling and semantic field strength driving Technical Field

[0001] This invention relates to the field of water conservancy knowledge graph updating technology, and more specifically, to a dynamic real-time updating system and method for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driving. Background Technology

[0002] Knowledge graphs, as structured semantic knowledge bases, are core infrastructure for smart water conservancy and have been explored and applied in scenarios such as joint scheduling of cascade reservoirs and watershed flood control decision-making. However, current research and application of water conservancy knowledge graphs suffer from the pain points of "building without innovation and using without flexibility." Specific defects and root causes are as follows: Passive and lagging update mode: Existing water conservancy knowledge graph updates rely on periodic ETL processes or manual input, which cannot match the dynamic characteristics of water conservancy systems. For example, if the outflow from a cascade reservoir changes drastically during the flood season, traditional graphs only update the data the next day, leading to delayed scheduling decisions. Existing stream processing technologies can only perform "data-data" conversions and cannot establish "physical state-semantic knowledge" associations, remaining a passive response.

[0003] The data and knowledge semantic gap is widening: multimodal data from water conservancy systems lacks semantic raw information and is disconnected from the semantic association of knowledge graphs. Existing technologies mostly use simple threshold triggers, but the impact of water conservancy physical quantities is continuous, and threshold mechanisms cannot quantify the semantic value of gradual changes, making graph updates unable to reflect the true state.

[0004] The current hydraulic knowledge graph lacks endogenous evolutionary momentum based on physical mechanisms. It is a "static information container," with updates relying on external program rules. Rule formulation depends on human experience and is disconnected from physical mechanisms. For example, the impact of river gate opening on downstream water levels cannot be addressed by existing rules that cannot embed physical parameters, leading to updates that do not reflect reality. Existing optimization techniques can only complete relationships on a static graph and cannot incorporate real-time states to achieve a "physical-information" closed loop.

[0005] Unified semantic fusion of multi-source heterogeneous data is challenging in water conservancy. Water conservancy data exhibits significant modal differences, including time-series, spatial, text, and image data. Existing technologies often employ a "divide and conquer" strategy (e.g., storing time-series data in a time-series database and processing text data using NLP tools), but lack a unified semantic framework to map multi-modal data into semantically understandable units in the graph, leading to fragmented graph updates. For example, if a reservoir collects data such as "water level 142.5m," "gate opening 40%," and "downstream flow velocity 3m / s," existing technologies cannot correlate this with the semantic knowledge of "reservoir discharge causing increased downstream flow velocity." Existing improvement schemes (such as CEP-based complex event processing) attempt to correlate multi-source data through "event rules," but these rules require manual pre-definition, cannot adapt to different water conservancy scenarios, and lack embedding of water conservancy physical mechanisms, resulting in insufficient universality and accuracy. Therefore, there is an urgent need in this field for a new knowledge graph updating technology that takes physical mechanisms as its core, semantic computing as its bridge, and dynamic evolution as its goal, to solve the pain point of "building without updating" and provide dynamic support for digital twin watersheds.

[0006] Therefore, it is necessary to design a dynamic real-time update system and method for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driving to solve the problems existing in the current technology. Summary of the Invention

[0007] In view of this, the present invention proposes a dynamic real-time update system and method for water conservancy knowledge graph based on deep physical-information coupling and semantic field strength driving, aiming to solve the above problems.

[0008] This invention proposes a dynamic real-time update system for a water conservancy knowledge graph based on deep physical-information coupling and semantic field strength driving, comprising: a physical perception layer for real-time acquisition, transmission, and preprocessing of multi-source water conservancy physical data; a coupling mapping layer for mapping multi-source water conservancy physical data into semantic state variables usable in the knowledge graph; a field strength calculation layer for calculating semantic field strength based on changes in semantic state variables and determining the nodes and relationships in the knowledge graph that need to be updated; a graph operation layer for executing knowledge graph update decisions and writing operations based on semantic field strength trigger thresholds; and a knowledge service layer for providing dynamic knowledge services to external smart water conservancy systems.

[0009] Furthermore, the physical sensing layer includes: a time-series sensor data acquisition unit for acquiring time-series data from radar level gauges, GNSS displacement sensors, and flow meters; an image acquisition unit for acquiring video image data from industrial cameras; a GIS spatial data acquisition unit for acquiring GIS spatial data, including riverbank vector data; a text data acquisition unit for acquiring text data, including scheduling instructions and inspection reports; a data transmission module for transmitting time-series data and video image data using the MQTT protocol and transmitting GIS spatial data and text data using the HTTP protocol; and a data preprocessing module for performing 3σ outlier removal and linear interpolation of missing data on the time-series data.

[0010] Furthermore, the coupling mapping layer includes: a state variable definition unit, used to generate a corresponding knowledge graph node NE for any hydraulic entity E, and represent the dynamic state of the hydraulic entity E as a set of state variables SE={s1,s2,...,sn}; and a multimodal mapping unit, used to transform the original physical data D of different modalities into standardized values ​​of state variables si; wherein, a single state variable si is defined as: si Where name represents the name of the state variable; value represents the value of the state variable; timestamp represents the data acquisition timestamp; and domain represents the physical constraint domain of the state variable.

[0011] Furthermore, when transforming the raw physical data D of different modes into standardized values ​​of state variables si, the process includes: defining a mapping function. The original physical data D of different modes is transformed into state variables si; the time series data mapping is represented by the following formula:

[0012] in, This represents the numerical value of the state variable si; Represents a time-series data mapping function; The sensor's raw reading is represented by: a represents the calibration coefficient determined by the sensor calibration report; b represents the offset coefficient; the video image data mapping is expressed by the following formula:

[0013] in, This represents the numerical value of the gate opening state variable; It represents the video image data mapping function; It represents the timestamp. Image frames; This indicates the visible height of the gate obtained through YOLOv8 segmentation; The total height of the gate is represented by the following formula:

[0014] in, This represents the numerical value of the state variable extracted from the text. This represents a text data mapping function; T represents text data; attr represents the attribute name to be extracted; This represents the attribute value extraction function; GIS spatial data mapping is represented by the following formula:

[0015] in, Indicates spatial semantic tags; This represents a GIS spatial data mapping function; G represents GIS vector data; Schema represents the predefined spatial semantics of the knowledge graph. This represents the tag matching function.

[0016] Furthermore, the field strength calculation layer includes: a semantic field strength calculation unit, used to calculate the semantic field strength of the state variable si for the target node NT; wherein, for the change of the state variable si, the semantic field strength generated by it for the target node NT in the knowledge graph is a combination of scalar intensity and direction; and a change significance calculation unit, used to calculate the change significance ξ(si) of the state variable based on the numerical change, the acceleration of change, and the deviation of the physical threshold of the state variable si. The semantic relevance calculation unit is used to calculate the semantic relevance ρ(NS,NT,siname) between the source node NS and the target node NT based on the predefined schema and semantic attributes of the state variables in the knowledge graph. .

[0017] Furthermore, the field strength calculation layer also includes: a spatiotemporal attenuation coefficient calculation unit, used to calculate the spatiotemporal attenuation coefficient γ(NS,NT) based on the shortest path length between the source node NS and the target node NT in the knowledge graph and a preset attenuation coefficient. Based on the calculation results, the semantic field strength Ψsi→NT of the state variable si to the target node NT is represented as a combination of scalar strength and direction, which is used to drive the dynamic update of the knowledge graph.

[0018] Furthermore, the graph operation layer includes: an operation triggering module, used to set a field strength triggering threshold Ψtthreshold, and trigger a knowledge graph update operation for the target node NT when the semantic field strength Ψsi→NT of the state variable si to the target node NT is greater than or equal to the field strength triggering threshold Ψtthreshold; an operation synthesis module, used to calculate the total field strength Ψtotal→NT of the target node according to the weight w(siname) of each state variable when the target node NT simultaneously receives semantic field strengths from multiple source nodes or state variables, and determine the direction with the largest total field strength as the final operation direction; and an operation execution module, used to convert the final operation direction into a specific update instruction for the knowledge graph node NT, and execute the corresponding update operation through the knowledge graph database.

[0019] Furthermore, when the operation execution module performs the corresponding update operation through the knowledge graph database, the update operation includes: UPDATE_ATTRIBUTE operation, used to update the attribute value and update time of the target node NT; ACTIVATE_RELATION operation, used to activate the semantic relationship between the source node NS and the target node NT; and CREATE_ALARM_RELATION operation, used to create an early warning relationship node between the source node NS and the target node NT to generate real-time water conservancy early warning information.

[0020] Furthermore, the knowledge service layer includes: an interface module for providing knowledge graph services to the outside world through standardized interfaces, including REST API and WebSocket; a knowledge query module for returning the current state information of the target water conservancy entity based on external call requests; a relationship reasoning module for reasoning the future state information of the target water conservancy entity based on upstream and downstream relationships and other semantic relationships in the knowledge graph; and an early warning generation module for generating real-time flood control early warning information based on flood risk relationships in the knowledge graph and pushing it to smart water conservancy applications through the standardized interfaces.

[0021] Compared with existing technologies, the beneficial effects of this invention are as follows: The dynamic real-time update system for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driving provided by this invention can effectively integrate multi-source water conservancy data. Through the physical perception layer, it achieves comprehensive acquisition and preprocessing of real-time data, ensuring data accuracy and integrity. The design of the coupling mapping layer enables the raw physical data to be efficiently transformed into semantic state variables required by the knowledge graph, thus providing basic support for subsequent dynamic updates. The field strength calculation layer, through precise calculation of semantic field strength, can quickly locate the knowledge graph nodes and relationships that need to be updated, significantly improving update efficiency. The graph operation layer, based on the semantic field strength triggering mechanism, realizes intelligent update decision-making and execution, ensuring the real-time performance and accuracy of the knowledge graph. The knowledge service layer provides flexible knowledge services to external systems through standardized interfaces, further enhancing the system's practicality and scalability.

[0022] In another aspect, this invention also proposes a real-time navigation method for orthopedic surgery based on multimodal image fusion and artificial intelligence, comprising the following steps: real-time acquisition, transmission, and preprocessing of multi-source hydraulic physical data; mapping the multi-source hydraulic physical data into semantic state variables usable in a knowledge graph; calculating the semantic field strength based on changes in the semantic state variables and determining the nodes and relationships in the knowledge graph that need to be updated; executing knowledge graph update decisions and writing operations according to the semantic field strength trigger threshold; and providing dynamic knowledge services to an external smart water conservancy system.

[0023] It is understandable that the above-mentioned dynamic real-time update system and method for water conservancy knowledge graph based on deep physical-information coupling and semantic field strength driving have the same beneficial effects, and will not be elaborated here. Attached Figure Description

[0024] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 is a structural block diagram of a dynamic real-time update system for a water conservancy knowledge graph based on deep physical-information coupling and semantic field strength driven by an embodiment of the present invention; Figure 2 is a flowchart of a dynamic real-time update method for a water conservancy knowledge graph based on deep physical-information coupling and semantic field strength driven by an embodiment of the present invention. Detailed Implementation

[0025] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the disclosure to those skilled in the art. It should be noted that, unless otherwise specified, embodiments and features in the embodiments of the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0026] Referring to Figure 1, in some embodiments of this application, this embodiment provides a dynamic real-time update system for a water conservancy knowledge graph based on deep physical-information coupling and semantic field strength driving, comprising: a physical perception layer for real-time acquisition, transmission, and preprocessing of multi-source water conservancy physical data; a coupling mapping layer for mapping multi-source water conservancy physical data into semantic state variables usable in the knowledge graph; a field strength calculation layer for calculating semantic field strength based on changes in semantic state variables and determining the nodes and relationships in the knowledge graph that need to be updated; a graph operation layer for executing knowledge graph update decisions and writing operations according to semantic field strength trigger thresholds; and a knowledge service layer for providing dynamic knowledge services to external smart water conservancy systems.

[0027] It is understood that the dynamic real-time update system for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driven by this embodiment can effectively integrate multi-source water conservancy data. Through the physical perception layer, it achieves comprehensive acquisition and preprocessing of real-time data, ensuring data accuracy and integrity. The design of the coupling mapping layer enables the raw physical data to be efficiently transformed into semantic state variables required by the knowledge graph, thus providing basic support for subsequent dynamic updates. The field strength calculation layer, through precise calculation of semantic field strength, can quickly locate the knowledge graph nodes and relationships that need to be updated, significantly improving update efficiency. The graph operation layer, based on the semantic field strength triggering mechanism, realizes intelligent update decision-making and execution, ensuring the real-time performance and accuracy of the knowledge graph. The knowledge service layer provides flexible knowledge services to external systems through standardized interfaces, further enhancing the system's practicality and scalability.

[0028] Specifically, the physical sensing layer includes: a time-series sensor data acquisition unit for acquiring time-series data from radar level gauges, GNSS displacement sensors, and flow meters; an image acquisition unit for acquiring video image data from industrial cameras; a GIS spatial data acquisition unit for acquiring GIS spatial data, including riverbank vector data; a text data acquisition unit for acquiring text data, including dispatch instructions and inspection reports; a data transmission module for transmitting time-series data and video image data using the MQTT protocol and transmitting GIS spatial data and text data using the HTTP protocol; and a data preprocessing module for performing 3σ outlier removal and linear interpolation of missing data on the time-series data.

[0029] Understandably, the design of the physical sensing layer fully considers the diversity and complexity of multi-source data, achieving comprehensive coverage of water conservancy-related data through the collaborative work of different units. The time-series sensor data acquisition unit efficiently acquires real-time data from devices such as radar level gauges, GNSS displacement sensors, and flow meters, providing high-precision dynamic information for subsequent analysis. The image acquisition unit focuses on video image data from industrial cameras, ensuring visualized monitoring of water conservancy projects. The GIS spatial data acquisition unit not only provides riverbank vector data but also supports the integration of other geographic information data, further enriching the system's spatial analysis capabilities. The text data acquisition unit, by acquiring scheduling instructions and inspection reports, combines human experience with automated systems, enhancing the comprehensive value of the data. The data transmission module selects appropriate protocols based on different data types, ensuring the efficiency and stability of data transmission. The data preprocessing module effectively improves data quality by performing 3σ outlier removal and linear interpolation of missing data on the time-series data, laying a solid foundation for subsequent knowledge graph construction.

[0030] Specifically, the coupling mapping layer includes: a state variable definition unit, used to generate a corresponding knowledge graph node NE for any hydraulic entity E, and represent the dynamic state of the hydraulic entity E as a set of state variables SE={s1,s2,...,sn}; and a multimodal mapping unit, used to transform the original physical data D of different modalities into standardized values ​​of state variables si; wherein, a single state variable si is defined as: si Where name represents the name of the state variable; value represents the value of the state variable; timestamp represents the data acquisition timestamp; and domain represents the physical constraint domain of the state variable.

[0031] Specifically, transforming the raw physical data D of different modes into standardized values ​​of state variables si includes: defining a mapping function. The original physical data D of different modes is transformed into state variables si; the time series data mapping is represented by the following formula:

[0032] in, This represents the numerical value of the state variable si; Represents a time-series data mapping function; The sensor's raw reading is represented by: a represents the calibration coefficient determined by the sensor calibration report; b represents the offset coefficient; the video image data mapping is expressed by the following formula:

[0033] in, This represents the numerical value of the gate opening state variable; It represents the video image data mapping function; It represents the timestamp. Image frames; This indicates the visible height of the gate obtained through YOLOv8 segmentation; The total height of the gate is represented by the following formula:

[0034] in, This represents the numerical value of the state variable extracted from the text. This represents a text data mapping function; T represents text data; attr represents the attribute name to be extracted; This represents the attribute value extraction function; GIS spatial data mapping is represented by the following formula:

[0035] in, Indicates spatial semantic tags; This represents a GIS spatial data mapping function; G represents GIS vector data; Schema represents the predefined spatial semantics of the knowledge graph. This represents the tag matching function.

[0036] Understandably, the design of the coupling mapping layer aims to achieve efficient transformation of multi-source water conservancy data into semantic state variables of the knowledge graph. The state variable definition unit generates corresponding knowledge graph nodes for each water conservancy entity and abstracts its dynamic state into a set of state variables, providing a clear structured framework for subsequent data processing and analysis. The multimodal mapping unit further refines this process, designing specialized mapping methods for the raw physical data of different modalities to ensure accurate standardization and representation of the data. Time-series data mapping improves the accuracy of converting raw readings into state variable values ​​by introducing sensor calibration coefficients and offset coefficients, thus ensuring the reliability of dynamic monitoring data. Video image data mapping utilizes advanced computer vision technologies, such as the YOLOv8 segmentation algorithm, to automate the calculation of key indicators such as gate opening, significantly reducing the need for manual intervention. Text data mapping leverages natural language processing technology to extract key attribute values ​​from scheduling instructions and inspection reports, transforming unstructured text into structured state variables and enhancing the system's intelligence level. GIS spatial data mapping connects geographic information with the spatial semantics of the knowledge graph through a label matching mechanism, providing the system with powerful spatial analysis capabilities. This multi-layered, multi-modal mapping mechanism not only improves the efficiency and accuracy of data transformation but also lays a solid foundation for semantic field strength analysis in the subsequent field strength calculation layer. Through the processing of the coupled mapping layer, the original physical data can be seamlessly integrated into the knowledge graph system, providing a reliable guarantee for the dynamic updating and real-time application of water conservancy knowledge.

[0037] Specifically, the field strength calculation layer includes: a semantic field strength calculation unit, used to calculate the semantic field strength of the state variable si for the target node NT; wherein, for the change of the state variable si, the semantic field strength generated by it for the target node NT in the knowledge graph is a combination of scalar intensity and direction; and a change significance calculation unit, used to calculate the change significance ξ(si) of the state variable based on the numerical change, acceleration, and physical threshold deviation of the state variable si. The semantic relevance calculation unit is used to calculate the semantic relevance ρ(NS,NT,siname) between the source node NS and the target node NT based on the predefined schema and semantic attributes of the state variables in the knowledge graph. .

[0038] In this embodiment, the concept of "field" in physics is used to transform the changes in the state of physical entities into the driving force for the evolution of knowledge graphs—semantic field strength. By quantifying the "intensity" and "direction" of the field strength, the update requirements of the graph are determined.

[0039] Semantic field strength definition: for state variables The changes in the target node within the knowledge graph The resulting semantic field strength is a combination of scalar intensity and direction, where the scalar intensity... The definition of represents

[0040] in, Represents state variables For the target node The semantic field strength scalar intensity (dimensionless, value range 0~100). Represents state variables Significance of change (dimensionless, measuring the degree of drastic change in physical state). Represents semantic relevance (dimensionless, measuring the source node) With the target node pass (the degree of closeness of the relationship); This represents the spatiotemporal attenuation coefficient (dimensionless, simulating the attenuation law of physical influences in space and time). Represents state variables The source node (i.e. the graph node corresponding to the hydraulic entity that generates this state, such as the "Shanxi Dam" node).

[0041] Calculation of the significance of change: The significance of change comprehensively considers the "rate of change", "acceleration of change" and "deviation from the physical threshold" of the physical quantity. The formula is as follows:

[0042] in, The weighting coefficient for the rate of change (determined based on the type of physical quantity, such as slowly changing physical quantities like water level) Rapidly changing physical quantities such as gate opening ); Represents state variables The numerical change ( Units and Consistent); This indicates the data acquisition time interval (unit: seconds, such as water level data). ); Indicates the weighting coefficient of the changing acceleration (a slowly varying physical quantity) Rapidly changing physical quantities ); Represents state variables The second difference ( Units and Consistent (s); This represents the state value function (dimensionless, measuring the degree to which a physical quantity deviates from a critical threshold).

[0043] Among them, the state value function Based on the threshold settings of water conservancy projects (taking reservoir water level as an example):

[0044] in, This indicates the warning threshold for water conservancy entities (such as the warning value of the Shanxi Dam water level). ); This represents the lower limit of the normal threshold for a water conservancy entity (e.g., the lower limit of the normal water level at the Shanxi Dam). ); Indicates the amplification factor for exceeding the warning threshold ( (Set according to SL601-2021 specification). This indicates an amplification factor below the normal threshold ( ).

[0045] Semantic relevance calculation: Semantic relevance is determined based on the predefined schema of the knowledge graph (such as "upstream and downstream relationships" and "inclusion relationships") and the semantic attributes of state variables. The formula is as follows:

[0046] in, Indicates the source node With the target node Direct relationships between them (such as "downstreamOf" or "hasGate"). Indicates a direct relationship A set of associated attributes (predefined, such as the "downstreamOf" relationship) ); Indicates the direct correlation attenuation coefficient ( Such as the "downstreamOf" relationship ); This represents the semantic path between the source node and the target node (e.g., "Shanxi Dam - (downstreamOf) -> Zhaoshan Ferry - (downstreamOf) -> XX segment"). The set of attributes representing semantic path associations (such as the path mentioned above) ); Indicates the indirect correlation attenuation coefficient ( , as in the path described above ); Indicates the minimum value ( (This indicates no association).

[0047] Specifically, the field strength calculation layer further includes a unit for calculating the spatiotemporal attenuation coefficient, used to calculate the spatiotemporal attenuation coefficient γ(NS,NT) based on the shortest path length between the source node NS and the target node NT in the knowledge graph and a preset attenuation coefficient. Based on the calculation results, the semantic field strength Ψsi→NT of the state variable si to the target node NT is represented as a combination of scalar strength and direction, which is used to drive the dynamic update of the knowledge graph.

[0048] In this embodiment, the spatiotemporal attenuation coefficient simulates the attenuation law of physical influence with "map path length" (spatial correlation), and adopts an exponential attenuation model:

[0049] in, This represents the attenuation coefficient (determined based on the connectivity of the water conservancy network, such as the cascade hubs in the basin). Small waterways ); Indicates the source node With the target node The shortest path length (number of edges) in a knowledge graph, such as from Shanxi Dam to Zhaoshan Ferry. ).

[0050] Semantic field strength direction determination: The field strength direction determines the specific update operation type of the graph, based on the joint determination of "state variable type, numerical change trend, and node type". The determination rules are as follows (taking a water conservancy scenario as an example): If and for If the attribute (such as "WaterLevel") is specified, the direction is "UPDATE_ATTRIBUTE" (update node attribute); if and ,but Direction is (Activation relationship); The direction is "CREATEALARMRELATION" (create alert relationship); , , (Freeze attributes to avoid repeated updates).

[0051] Specifically, the graph operation layer includes: an operation triggering module, used to set a field strength triggering threshold Ψtthreshold, and trigger a knowledge graph update operation for the target node NT when the semantic field strength Ψsi→NT of the state variable si to the target node NT is greater than or equal to the field strength triggering threshold Ψtthreshold; an operation synthesis module, used to calculate the total field strength Ψtotal→NT of the target node according to the weight w(siname) of each state variable when the target node NT simultaneously receives semantic field strengths from multiple source nodes or state variables, and determine the direction with the largest total field strength as the final operation direction; and an operation execution module, used to convert the final operation direction into a specific update instruction for the knowledge graph node NT, and execute the corresponding update operation through the knowledge graph database.

[0052] In this embodiment, the knowledge graph operation layer receives the field strength calculation results and achieves accurate updates to the knowledge graph through a three-step process of "threshold triggering - operation synthesis - instruction execution": the operation triggering condition sets the field strength trigger threshold. (Adaptively adjusted based on node importance: core nodes such as reservoirs) Secondary nodes such as monitoring stations ),when At that time, triggering the target The update operation is represented by the following formula:

[0053] in, Indicates the field strength trigger threshold (dimensionless). Indicates the direction of the semantic field strength (i.e., the type of update operation); Indicates the operation trigger function (generated for) (Update instructions).

[0054] Operation synthesis when target node When receiving field strengths from multiple source nodes / state variables simultaneously, operation synthesis is required. The direction with the largest total field strength is taken as the final operation direction. The formula for the total field strength is:

[0055] in, Represents the target node Total field strength (dimensionless); The weights of the state variables are indicated (set according to the importance of water resources, such as the weight of "WaterLevel"). "GateOpening" weight ); Indicates that all actions applied to State variables and source node Sum.

[0056] Only when is the final operation performed, and the operation direction is the dominant direction corresponding to the total field strength (i.e., the direction of the field strength with the largest contribution).

[0057] The operation is executed using the knowledge graph database Neo4j 5.10 (supporting high-concurrency writing, with a write throughput of ≥1000 TPS) as the storage engine, and the update operation is executed through Cypher statements. The examples are as follows: UPDATE_ATTRIBUTE: MATCH(n:Reservoir{id:"Shanxi Dam"}) SET n.GateOpening = 30, n.updateTime = "YYYY-08-01 10:00:00"; ACTIVATE_RELATION: MATCH(a:Reservoir{id:"Shanxi Dam"})-[r:downstreamOf]->(b:Dam{id:"Zhaoshandu"}) SET r.isActive = true; CREATE_ALARM_RELATION: MATCH(a:Reservoir{id:"Shanxi Dam"}),(b:FloodRiskArea{id:"XX section"}) CREATE(a)-[r:hasFloodRisk{level:"high",createTime:"YYYY-08-01 10:05:00"}]->(b).

[0058] It can be understood that in practical applications, this system can significantly improve the dynamic response ability of the water conservancy knowledge graph. By combining physical models with information processing technologies, accurate modeling and real-time updating of complex water conservancy scenarios are achieved. For example, in the management of cascade hubs in a basin, the system can automatically adjust the node attributes according to the water level changes upstream and downstream, ensuring the timeliness and accuracy of the data. At the same time, the direction determination mechanism of the semantic field strength provides clear guidance for the operations of the knowledge graph, avoiding unnecessary redundant updates and improving the operating efficiency of the system. In addition, the threshold setting of the operation trigger module further enhances the robustness of the system, enabling it to operate stably under noise interference. This multi-level and multi-dimensional design concept makes this method perform excellently in dealing with large-scale water conservancy networks and provides strong support for the construction of smart water conservancy.

[0059] Specifically, when the operation execution module performs the corresponding update operation through the knowledge graph database, the update operation includes: UPDATE_ATTRIBUTE operation, used to update the attribute value and update time of the target node NT; ACTIVATE_RELATION operation, used to activate the semantic relationship between the source node NS and the target node NT; and CREATE_ALARM_RELATION operation, used to create an early warning relationship node between the source node NS and the target node NT to generate real-time water conservancy early warning information.

[0060] Understandably, the execution of these operations relies on the system's real-time parsing and semantic understanding of multi-source data in water conservancy scenarios. By dynamically adjusting node attributes and relationships, the system can quickly respond to changes in the water conservancy network, thereby providing more accurate support for decision-making. For example, in flood warning scenarios, the CREATE_ALARM_RELATION operation can combine historical data with real-time monitoring information to generate warning nodes with high confidence, helping relevant departments to formulate response measures in advance. At the same time, the ACTIVATE_RELATION operation can strengthen the correlation between key nodes, ensuring the efficiency and accuracy of information transmission in emergency situations. This refined operational design not only enhances the practicality of knowledge graphs but also lays the foundation for the intelligent management of complex water conservancy systems.

[0061] Specifically, the knowledge service layer includes: an interface module for providing knowledge graph services to the outside world through standardized interfaces, including REST API and WebSocket; a knowledge query module for returning the current state information of the target water conservancy entity based on external call requests; a relationship reasoning module for reasoning the future state information of the target water conservancy entity based on upstream and downstream relationships and other semantic relationships in the knowledge graph; and an early warning generation module for generating real-time flood control early warning information based on flood risk relationships in the knowledge graph and pushing it to smart water conservancy applications through the standardized interfaces.

[0062] Understandably, the knowledge service layer is designed to provide comprehensive and efficient support for smart water management applications. The standardized design of the interface modules ensures seamless integration between the system and external applications, enabling stable operation for both real-time data transmission and batch query requests. The knowledge query module meets users' real-time monitoring needs by accurately reflecting the status information of target water entities. The relationship reasoning module further leverages the potential of knowledge graphs, predicting possible future state changes of target entities by analyzing upstream and downstream relationships and complex semantic networks, thus providing forward-looking suggestions for decision-makers. The early warning generation module plays a crucial role in emergency scenarios such as flood risks. It not only quickly generates early warning information but also enables real-time push notifications via WebSocket, ensuring relevant departments can obtain critical intelligence and take action immediately. This multi-layered service architecture significantly enhances the system's flexibility and adaptability, enabling it to cope with complex and ever-changing water management needs.

[0063] In this embodiment, the implementation steps of the present invention are described in detail using the "real-time update of the water conservancy knowledge graph of the Shanxi-Zhaoshandu cascade hub" as a specific implementation case.

[0064] (I) Implementation Preparation 1. Initialization of Physical Entities and Knowledge Graph Physical Entities: Shanxi Dam (Reservoir type), Zhaoshan Ferry (Dam type), XX Section (RiverReach type), Shanxi No. 1 Gate (Gate type); Initial State of Knowledge Graph: Nodes:

[0065] ;

[0066]

[0067] relation: ; .

[0068] 2. Equipment and Parameter Configuration: Sensor: RD-900 radar level gauge deployed at Shanxi Dam ( , ), gate opening monitoring camera (YOLOv8 model, Model parameters: (Gate opening is a rapidly changing physical quantity). , , , , , , .

[0069] (II) Implementation Steps Step 1: Physical Data Acquisition and Preprocessing (Time) Gate control system data: Raw data on the opening degree of Shanxi No. 1 gate (No systematic error, therefore) , (Water level sensor data: Shanxi Dam water level) (No change); Preprocessing: Data with no outliers or missing values ​​is directly fed into the coupling mapping layer.

[0070] Step 2: Coupling Mapping (Generating State Variables) Gate Opening Mapping: Using an image mapping function Segmentation ,but:

[0071] Generate state variables:

[0072] Water level mapping: (The value remains unchanged and will not be included in subsequent field strength calculations).

[0073] Step 3: Semantic field strength calculation (using (As the core state variable) 1. Calculate the significance of the change

[0074] , ; Rate of change term: ; Change in acceleration term: (For simplified calculation, the second-order difference is approximated as the first-order difference); State value function: Therefore ;final: (This is a simplified calculation; in actual engineering, due to sudden changes in gate opening, ...) The value is approximately 18, but the specific value needs to be scaled according to the actual parameters.

[0075] 2. Calculate semantic relevance.

[0076] right (Source node itself): ;right (Direct downstream): And "GateOpening" Therefore ;right (Indirect downstream): And "GateOpening" Therefore .

[0077] 3. Calculate the spatiotemporal decay coefficient

[0078] right : Therefore ;right : Therefore ;right : Therefore .

[0079] 4. Calculate semantic field strength

[0080] right : (Simplified value, actual engineering value approximately 18); For : (Actual project value approximately 8.73); For : (Actual engineering value is approximately 3.31).

[0081] 5. Determine the direction of the electric field strength. : Furthermore, "GateOpening" is an attribute, and its direction is "UPDATE_ATTRIBUTE"; for : and Direction="ACTIVATE_RELATION"; for Changes to "GateOpening" affect downstream flow, direction = "UPDATE_ATTRIBUTE" (marked "expectedFlowChange=positive").

[0082] Step 4: Triggering judgment for graph operation execution: In actual engineering All values ​​≥ 2.0, trigger operation; Operation execution: 1. Update : MATCH (n:Reservoir {id: "Shanxi Dam"}) SET n.GateOpening =30%, n.updateTime = "YYYY-08-01 10:00:00"; 2. Activate : MATCH (a:Reservoir {id: "Shanxi Dam"})-[r:downstreamOf]->(b:Dam {id: "Zhaoshan Ferry"}) SET r.isActive = true, r.activateTime = "YYYY-08-01 10:00:00";3. Update :MATCH (n:RiverReach {id: "XX segment"}) SET n.expectedFlowChange= "positive", n.updateTime = "YYYY-08-01 10:00:00".

[0083] Step 5: Subsequent Evolution (Time) Physical data: The water level sensor at Zhaoshandu detected that the water level had risen to 141.2m. The flow rate increased to 60m. 3 / s; Field strength calculation: with As the source node, with state variables "WaterLevel" and "FlowRate", calculate the... Field strength: Direction = "UPDATE_ATTRIBUTE"; Operation execution: MATCH (n:RiverReach {id: "XX segment"}) SET n.FlowRate = 60m 3 / s,n.expectedFlowChange = null, n.updateTime = "YYYY-08-01 10:05:00".

[0084] Referring to Figure 2, in some embodiments of this application, this embodiment provides a method for dynamic real-time updating of a water conservancy knowledge graph based on deep physical-information coupling and semantic field strength driving, including the following steps: S100: Real-time acquisition, transmission, and preprocessing of multi-source water conservancy physical data; S200: Mapping multi-source water conservancy physical data into semantic state variables usable in the knowledge graph; S300: Calculating the semantic field strength based on changes in the semantic state variables and determining the nodes and relationships in the knowledge graph that need to be updated; S400: Executing knowledge graph update decisions and writing operations according to the semantic field strength trigger threshold; S500: Providing dynamic knowledge services to external smart water conservancy systems.

[0085] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program goods. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program goods embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0086] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program goods according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0087] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0088] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0089] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A dynamic real-time update system for a water conservancy knowledge graph based on deep physical-information coupling and semantic field strength driving, characterized in that, include: The physical sensing layer is used for real-time acquisition, transmission, and preprocessing of multi-source hydraulic physical data; The coupling mapping layer is used to map multi-source hydraulic physics data into semantic state variables that can be used in the knowledge graph; the field strength calculation layer is used to calculate the semantic field strength based on the changes in the semantic state variables and determine the nodes and relationships that need to be updated in the knowledge graph; the graph operation layer is used to execute knowledge graph update decisions and write operations based on the semantic field strength trigger threshold. The knowledge service layer is used to provide dynamic knowledge services to external smart water conservancy systems.

2. The dynamic real-time update system for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driven according to claim 1, characterized in that, The physical sensing layer includes: a time-series sensor data acquisition unit for acquiring time-series data from radar level gauges, GNSS displacement sensors, and flow meters; an image acquisition unit for acquiring video image data from industrial cameras; a GIS spatial data acquisition unit for acquiring GIS spatial data, including riverbank vector data; a text data acquisition unit for acquiring text data, including dispatch instructions and inspection reports; a data transmission module for transmitting time-series data and video image data using the MQTT protocol and transmitting GIS spatial data and text data using the HTTP protocol; and a data preprocessing module for performing 3σ outlier removal and linear interpolation of missing data on the time-series data.

3. The dynamic real-time update system for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driving according to claim 2, characterized in that, The coupling mapping layer includes: a state variable definition unit, used to generate a corresponding knowledge graph node NE for any hydraulic entity E, and to represent the dynamic state of the hydraulic entity E as a set of state variables SE={s1,s2,...,sn}; and a multimodal mapping unit, used to transform the original physical data D of different modalities into standardized values ​​of state variables si; wherein, a single state variable si is defined as: si Where name represents the name of the state variable; value represents the value of the state variable; timestamp represents the data acquisition timestamp; and domain represents the physical constraint domain of the state variable.

4. The dynamic real-time update system for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driven according to claim 3, characterized in that, When transforming the raw physical data D of different modes into standardized values ​​of state variables si, the process includes: defining a mapping function. The original physical data D of different modes is transformed into state variables si; the time series data mapping is represented by the following formula: in, This represents the numerical value of the state variable si; Represents a time-series data mapping function; The sensor's raw reading is represented by: a represents the calibration coefficient determined by the sensor calibration report; b represents the offset coefficient; the video image data mapping is expressed by the following formula: in, This represents the numerical value of the gate opening state variable; It represents the video image data mapping function; It represents the timestamp. Image frames; This indicates the visible height of the gate obtained through YOLOv8 segmentation; The total height of the gate is represented by the following formula: in, This represents the numerical value of the state variable extracted from the text. This represents a text data mapping function; T represents text data; attr represents the attribute name to be extracted; This represents the attribute value extraction function; GIS spatial data mapping is represented by the following formula: in, Indicates spatial semantic tags; This represents a GIS spatial data mapping function; G represents GIS vector data; Schema represents the predefined spatial semantics of the knowledge graph. This represents the tag matching function.

5. The dynamic real-time update system for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driving according to claim 4, characterized in that, The field strength calculation layer includes: a semantic field strength calculation unit, used to calculate the semantic field strength of the state variable si for the target node NT; wherein, for the change of the state variable si, the semantic field strength generated by it for the target node NT in the knowledge graph is a combination of scalar intensity and direction; and a change significance calculation unit, used to calculate the change significance ξ(si) of the state variable based on the numerical change, acceleration, and physical threshold deviation of the state variable si. The semantic relevance calculation unit is used to calculate the semantic relevance ρ(NS,NT,siname) between the source node NS and the target node NT based on the predefined schema and semantic attributes of the state variables in the knowledge graph. 。 6. The dynamic real-time update system for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driving according to claim 5, characterized in that, The field strength calculation layer further includes a spatiotemporal attenuation coefficient calculation unit, used to calculate the spatiotemporal attenuation coefficient γ(NS,NT) based on the shortest path length between the source node NS and the target node NT in the knowledge graph and a preset attenuation coefficient. Based on the calculation results, the semantic field strength Ψsi→NT of the state variable si to the target node NT is represented as a combination of scalar strength and direction, which is used to drive the dynamic update of the knowledge graph.

7. The dynamic real-time update system for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driving according to claim 6, characterized in that, The graph operation layer includes: an operation triggering module, used to set a field strength triggering threshold Ψtthreshold, and trigger a knowledge graph update operation for the target node NT when the semantic field strength Ψsi→NT of the state variable si to the target node NT is greater than or equal to the field strength triggering threshold Ψtthreshold; an operation synthesis module, used to calculate the total field strength Ψtotal→NT of the target node according to the weight w(siname) of each state variable when the target node NT simultaneously receives the semantic field strength of multiple source nodes or state variables, and determine the direction with the largest total field strength as the final operation direction; and an operation execution module, used to convert the final operation direction into a specific update instruction for the knowledge graph node NT, and execute the corresponding update operation through the knowledge graph database.

8. The dynamic real-time update system for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driving according to claim 7, characterized in that, When the operation execution module performs the corresponding update operation through the knowledge graph database, the update operation includes: UPDATE_ATTRIBUTE operation, used to update the attribute value and update time of the target node NT; ACTIVATE_RELATION operation, used to activate the semantic relationship between the source node NS and the target node NT; and CREATE_ALARM_RELATION operation, used to create an early warning relationship node between the source node NS and the target node NT to generate real-time water conservancy early warning information.

9. The dynamic real-time update system for water conservancy knowledge graphs based on deep physical-information coupling and semantic field strength driving according to claim 8, characterized in that, The knowledge service layer includes: an interface module for providing knowledge graph services to the outside world through standardized interfaces, including RESTAPI and WebSocket; a knowledge query module for returning the current status information of the target water conservancy entity based on external call requests; a relationship reasoning module for reasoning the future status information of the target water conservancy entity based on upstream and downstream relationships and other semantic relationships in the knowledge graph; and an early warning generation module for generating real-time flood control early warning information based on flood risk relationships in the knowledge graph and pushing it to smart water conservancy applications through the standardized interfaces.

10. A method for dynamic real-time updating of a water conservancy knowledge graph based on deep physical-information coupling and semantic field strength driving, applied to the dynamic real-time updating system of a water conservancy knowledge graph based on deep physical-information coupling and semantic field strength driving as described in any one of claims 1-9, characterized in that, include: Real-time acquisition, transmission, and preprocessing of multi-source hydraulic physical data; Mapping multi-source hydraulic physics data into semantic state variables usable in knowledge graphs; Calculate the semantic field strength based on changes in semantic state variables and determine the nodes and relationships in the knowledge graph that need to be updated; execute knowledge graph update decisions and write operations based on the semantic field strength trigger threshold; and provide dynamic knowledge services to external smart water conservancy systems.

Images